A urinary control management system and method based on an intelligent urinary catheter

Abstract

The present invention relates to the technical field of healthcare informatics, and particularly relates to a urinary control management system and method based on an intelligent catheter, including: obtaining historical patient data during the catheter insertion process, thereby obtaining a scalar value sequence and a weight vector sequence of nodes in the hidden layer of the neural network architecture to screen out the target hidden layer, dividing each target hidden layer into several node groups, obtaining updated node groups, thereby obtaining a neural network with an updated architecture, training using a training data set, obtaining the trained neural network, thereby outputting the pain points during the catheter insertion process of the current patient, and performing local anesthesia on the pain points. The present invention can accurately identify the pain points during the catheter insertion process of the patient, and then perform local anesthesia on the pain points to improve the urinary control management effect.

Description

Technical Field

[0001] The present invention relates to the technical field of healthcare informatics, and particularly relates to a urinary control management system and method based on an intelligent catheter. Background Art

[0002] In urological surgeries, catheters are an essential medical device, mainly used for urine drainage management, bladder decompression, and urine volume monitoring. However, existing non-intelligent catheters have many problems in practical applications. For example, traditional catheters are usually made of hard materials (such as rubber or PVC), and they can easily cause mechanical irritation to the urethral mucosa during insertion. Especially in male patients, due to the longer and more curved urethra, the process of inserting a catheter is often accompanied by severe pain and discomfort, and it is impossible to effectively judge the physical changes of the patient during catheterization.

[0003] Existing problems: During the urinary control management of an intelligent catheter, microelectrodes on the catheter surface are used to record the urethral electrical signals during catheter insertion, and then a neural network model is used to analyze the urethral electrical signals to locate local discomfort (pain) points in the urethra, so as to perform local anesthesia on the pain points. However, when identifying the pain point coordinates in the urethra through the neural network, during catheter insertion, if there is pain at a certain position in the urethra, it means that there is irritation or damage to the urethra or surrounding tissues in this area, and these damaged areas will affect adjacent areas, and then the electrical signals in adjacent areas will also change. Therefore, each local discomfort point may affect the electrical signals in adjacent areas, resulting in the neural network learning more duplicate information, and thus it is difficult to accurately locate the pain point area. Summary of the Invention

[0004] The present invention provides a urinary control management system and method based on an intelligent catheter to solve the existing problems.

[0005] A urinary control management system and method based on an intelligent catheter of the present invention adopts the following technical solutions:

[0006] An embodiment of the present invention provides a urinary control management method based on an intelligent catheter, and the method includes the following steps:

[0007] Obtain a historical data set composed of all data of all historical patients during catheter insertion, and divide the historical data set into a test data set and a training data set; input each piece of data in the test data set into the neural network one by one, and output the scalar value sequence and weight vector sequence of each node in each hidden layer in the neural network architecture;

[0008] According to the magnitudes of the weight vectors in the weight vector sequence of each node in each hidden layer, several target hidden layers are selected; in each target hidden layer, all nodes are divided into several node groups according to the similarity between the scalar value sequences of the nodes;

[0009] The nodes in the node groups are pruned to obtain updated node groups;

[0010] In the neural network architecture, each updated node group of each target hidden layer is regarded as a node to obtain the neural network after updating the architecture; the neural network after updating the structure is trained using the training data set to obtain the trained neural network; the pain points during the catheter insertion process of the current patient are obtained using the trained neural network, and local anesthesia is performed on the pain points.

[0011] Further, the steps of selecting several target hidden layers specifically include the following:

[0012] Obtain the magnitude of each weight vector in the weight vector sequence of any node in any hidden layer to form a weight vector magnitude sequence;

[0013] According to the magnitudes of the weight vectors in the weight vector magnitude sequence of each node in the hidden layer, target nodes are selected;

[0014] According to the differences between adjacent weight vector magnitudes in the weight vector magnitude sequence of each target node, duplicate information nodes are selected;

[0015] According to the number of duplicate information nodes in each hidden layer, target hidden layers are selected.

[0016] Further, the steps of selecting target nodes according to the magnitudes of the weight vectors in the weight vector magnitude sequence of each node in the hidden layer specifically include the following:

[0017] Select any node A in any hidden layer, and use the median in the weight vector magnitude sequence of A as the reference weight value; obtain the number S of weight vector magnitudes in the weight vector magnitude sequence of A that are less than the reference weight value. When the ratio of the number S to the length of the weight vector magnitude sequence of A is greater than the preset first proportion threshold, A is recorded as a target node.

[0018] Further, the steps of selecting duplicate information nodes according to the differences between adjacent weight vector magnitudes in the weight vector magnitude sequence of each target node specifically include the following:

[0019] In the modulus sequence of the weight vectors of any target node B, select any two adjacent modulus values of the weight vectors C and D, and take the ratio of the absolute value of the difference between C and D to the maximum value of C and D as the difference value between C and D. When the maximum value among the difference values of all adjacent modulus values of the weight vectors is less than a preset judgment threshold, mark B as a repeated information node.

[0020] Further, the step of screening out the target hidden layer according to the number of repeated information nodes in each hidden layer includes the following specific steps:

[0021] In any hidden layer, when the ratio of the number of repeated information nodes to the number of nodes in the any hidden layer is less than a preset second proportion threshold, mark the any hidden layer as the target hidden layer.

[0022] Further, the step of dividing all nodes into several node groups according to the similarity between the scalar value sequences of the nodes in each target hidden layer includes the following specific steps:

[0023] In any target hidden layer, obtain the inverse ratio value of the cosine similarity between the scalar value sequences of any two nodes as the clustering distance between the any two nodes, and perform a clustering operation on all nodes to obtain several node groups.

[0024] Further, the step of deleting the nodes in the node group to obtain an updated node group includes the following specific steps:

[0025] Obtain the modulus of each weight vector in the weight vector sequence of any node in any hidden layer to form a modulus sequence of the weight vectors;

[0026] In the modulus sequence of the weight vectors of any node in any node group, obtain the mean value of all the modulus values of the weight vectors, and count the ordinal values of all the modulus values of the weight vectors greater than the mean value to form a set of weight ordinal values;

[0027] Obtain the input serial numbers of each piece of data when all pieces of data in the test dataset are input into the neural network one by one, and count the set of input serial numbers of all pieces of data of each historical patient in the test dataset;

[0028] According to the set of weight ordinal values and the set of input serial numbers of each historical patient, screen out the initial reserved nodes from any node group, and obtain the historical patients corresponding to each initial reserved node;

[0029] According to the historical patients corresponding to the initial reserved nodes, screen out the final reserved nodes from the initial reserved nodes;

[0030] Form an updated node group with all the final reserved nodes in any node group.

[0031] Further, the steps of screening out initial retained nodes from any node group according to the set of weighted order values and the set of input serial numbers of each historical patient, and obtaining the historical patient corresponding to each initial retained node are as follows:

[0032] If the set of weighted order values of any node in any node group is a subset of the set of input serial numbers of any historical patient, then mark the any node as an initial retained node, and use the any historical patient as the historical patient corresponding to the initial retained node.

[0033] Further, the steps of screening out final retained nodes from the initial retained nodes according to the historical patients corresponding to the initial retained nodes are as follows:

[0034] In any node group, use the historical patients corresponding to all the initial retained nodes to form a historical patient set, and mark the initial retained node corresponding to the historical patient with the largest number of the same historical patients in the historical patient set as the final retained node.

[0035] The present invention also provides a urinary control management system based on an intelligent catheter, including a memory, a processor, and a computer program stored on the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the foregoing urinary control management method based on an intelligent catheter.

[0036] The beneficial effects of the technical solution of the present invention are:

[0037] In the embodiment of the present invention, a historical data set composed of historical patient data during the catheter insertion process is obtained, so as to output a scalar value sequence and a weight vector sequence of each node in each hidden layer of the neural network architecture, thereby screening out several target hidden layers, dividing each target hidden layer into several node groups, deleting the nodes in the node groups to obtain updated node groups, and using each updated node group as a node. Thus, the influence caused by the electrical signals in the adjacent areas of the pain points in the urethra is reduced, and further, the neural network is prevented from learning more repetitive information, so that the pain point area can be accurately located, and a neural network after updating the architecture is obtained, which is trained using a training data set to obtain a trained neural network. It is used to output the pain points during the catheter insertion process of the current patient, and local anesthesia is performed on the pain points. Thus, the present invention can accurately identify the pain points during the catheter insertion process of the patient, and then perform local anesthesia on the pain points to improve the urinary control management effect. Description of the Drawings

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0039] Figure 1 It is a flowchart of the steps of a urine control management method based on an intelligent catheter of the present invention;

[0040] Figure 2 It is a schematic diagram of catheter insertion;

[0041] Figure 3 It is a schematic diagram of a neural network architecture;

[0042] Figure 4 It is a schematic diagram of the neural network after updating the architecture. Specific Embodiments

[0043] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific embodiments, structures, features, and effects of a urine control management system and method based on an intelligent catheter proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0045] The following specifically describes the specific solutions of a urine control management system and method based on an intelligent catheter provided by the present invention in conjunction with the accompanying drawings.

[0046] Please refer to Figure 1 , which shows a flowchart of the steps of a urine control management method based on an intelligent catheter provided by an embodiment of the present invention. The method includes the following steps:

[0047] Step S001: Obtain the historical dataset composed of all the data of all historical patients during the catheter insertion process, and divide the historical dataset into a test dataset and a training dataset; input each piece of data in the test dataset into the neural network one by one, and output the scalar value sequence and weight vector sequence of each node in each hidden layer in the neural network architecture.

[0048] It should be noted that in this embodiment, the urinary catheter adopts a double-layer design. The inner layer conducts urine drainage. It is made of biocompatible materials to ensure that urine can flow out of the body smoothly and will not cause harm to the urethra or bladder. Moreover, the inner layer material has high flexibility and compressive resistance to adapt to the physiological curve of the urethra. The interlayer is configured with anesthetic components such as lidocaine, and lidocaine is slowly released through the permeation port. The function of the interlayer is to provide anesthetic effect on the local area in the urethra, especially to relieve discomfort or pain during the insertion of the urinary catheter. The surface layer is the part of the urinary catheter with microelectrodes. The microelectrodes collect the electrophysiological signals of the urethra by contacting the inner wall of the urethra on the surface. The schematic diagram of the urinary catheter insertion is as Figure 2 shown. Thus, the surface microelectrodes are used to record the urethral electrical signals during the insertion of the urinary catheter, and then the neural network model analyzes the urethral electrical signals to locate the local discomfort (pain) points in the urethra. Finally, the local anesthetic in the urethra is administered in the form of local drug delivery through the permeation port on the surface layer.

[0049] For several historical patients, the voltage and three-dimensional coordinates of each microelectrode on the surface of the urinary catheter of each historical patient at each moment during the insertion of the urinary catheter are collected.

[0050] It should be noted that in this embodiment, the data acquisition frequency is once per second. The microelectrodes configured on the surface layer of the urinary catheter adopt nanoscale electrode materials, which have high sensitivity and can detect weak electrophysiological signals in the urethra. The electrodes contact the urethral tissue through the tiny orifices on the surface layer to collect the electrical signals in the urethra. During the insertion of the urinary catheter, the microelectrodes collect the electrical signals of the urethra in real time. These signals mainly come from the physiological activities of the urethral wall, such as the muscle contraction and potential change of the urethra. During the insertion process, due to the stress response of the urethral tissue, local potential changes or pain signals may be generated, and these signals can be captured by the microelectrode sensors. Then, the three-dimensional coordinates of the microelectrodes are measured by a coordinate measuring machine through contact or non-contact methods. This is a high-precision measuring device.

[0051] The neural network in this embodiment adopts a CNN (Convolutional Neural Network) deep network model and selects the ResNet (Residual Network) architecture. Among them, the ResNet architecture consists of an input layer, an output layer, and several hidden layers.

[0052] The input of the neural network is as follows: For each microelectrode on the surface of the catheter of each historical patient during the catheter insertion process, a piece of data is formed by three voltages within every three seconds and the three-dimensional coordinates in the last second (e.g., the first piece of data includes the voltages at the 1st second, 2nd second, and 3rd second and the three-dimensional coordinates at the 3rd second, and the second piece of data includes the voltages at the 4th second, 5th second, and 6th second and the three-dimensional coordinates at the 6th second. If the last period does not meet 3 seconds, it is not regarded as a piece of data). Taking this as an example for description, in other embodiments, it can be set as other acquisition methods for each piece of data, which is not limited in this embodiment. All pieces of data of all historical patients form a historical dataset, which serves as the input of the neural network.

[0053] It should be noted that: This ensures that the input lengths of the network at different times are consistent, avoiding the length of the electrical signal sequence becoming longer and longer over time, resulting in different lengths of electrical signal sequences at different moments, and thus being unable to be input into the same neural network.

[0054] Among them, each piece of data contains 6 features, namely the voltages for 3 consecutive seconds and the coordinate values in three dimensions. Therefore, the number of input nodes of the neural network is set to 6.

[0055] The output of the neural network is: the three-dimensional coordinates of the pain point.

[0056] Schematic diagram of the neural network architecture, as Figure 3 shown, Figure 3 Each circle in it represents a node in the neural network. The nodes in the first column from left to right are the input layer, the nodes in the last column are the output layer, and the nodes in each other column are each hidden layer.

[0057] In this embodiment, in the historical dataset, according to the moment when each historical patient reflects pain, each piece of data corresponding to the moment of pain of each historical patient is obtained as the electrical signal data of the pain generated by the historical patient during catheter insertion. And doctors or professional medical staff label the data according to the patient's symptoms, physiological signals, and clinical diagnosis to mark which piece of data corresponds to the pain signal, so as to guide the neural network model to learn how to distinguish pain and non-pain signals.

[0058] It should be noted that: Due to the certain diffusibility of nerve conduction, pain signals will cause corresponding signal changes in adjacent areas, that is, the pain signals during catheter insertion will spread between different areas through nerve conduction, especially for relatively extensive or deep pain stimuli. This will make the neural network contain more repetitive information, and thus make it difficult for the neural network to accurately distinguish the pain at a specific location. To this end, in this embodiment, through the method of grouped convolution, the hidden layer nodes containing more repetitive information are grouped together, and then in each hidden layer, the hidden nodes containing more repetitive information are given a smaller proportion of the allowable error.

[0059] In the historical dataset, randomly select data entries to form a test dataset, where is the total number of all data entries in the historical dataset, is the floor function. Use all the data entries outside the test dataset to form a training dataset. This is used as an example for description, and in other implementation manners, it can be set to other allocation methods, which are not limited in this embodiment. Thus, based on the data during the training process of the test dataset and the data obtained after the training ends, neuron nodes in certain hidden layers of the neural network are grouped.

[0060] During the neural network training process of the test dataset, when the first data entry is input into the network, each node in each hidden layer will obtain a scalar value, and the weight value of each edge in the neural network is obtained through the feedback of the loss value calculated by the loss function in the neural network. The weight values of all edges corresponding to each node in each hidden layer form a weight vector. Then, when the second data entry is input into the network, each node in the hidden layer obtains an updated scalar value, and the updated weight value of each edge in the neural network is obtained through the feedback of the loss value calculated by the loss function in the neural network. This is iteratively updated in sequence until all data entries in the test dataset are input. Among them, the loss value calculated by the loss function in the neural network measures the difference between the predicted output of the model and the true label. When there is a difference, the edge weights in the network are updated based on the difference.

[0061] Thus, after all data entries in the test dataset are input into the neural network one by one, a scalar value sequence and a weight vector sequence of each node in each hidden layer in the neural network architecture are obtained.

[0062] It should be noted that: each scalar value in the scalar value sequence corresponds to one data entry in the test dataset, each weight vector in the weight vector sequence corresponds to one data entry in the test dataset. The scalar value sequence represents the change process of the value of each node in the hidden layer during the neural network training process, and the weight vector sequence represents the process of iterative update of the weights of each node in the hidden layer during the neural network training process. The above explanations of the neural network are all well-known technologies.

[0063] Step S002: According to the magnitudes of the norms of the weight vectors in the weight vector sequence of the nodes in each hidden layer, select several target hidden layers; in each target hidden layer, divide all nodes into several node groups according to the similarity between the scalar value sequences of the nodes.

[0064] It should be noted that: if in a certain hidden layer, the scalar value sequences of several nodes are similar, indicating that the changes of these nodes during pain generation in different patients are similar, then these nodes may represent more repetitive information that is transmitted from the urethral nerves where the pain points are located to adjacent nerves, resulting in pain stimuli also occurring at the adjacent locations. For all hidden layers, not every hidden layer can concentrate repetitive information into several nodes. It is possible that each node in a certain hidden layer contains some repetitive information. In this case, it is difficult to process the repetitive information through grouping. Therefore, it is necessary to first find the hidden layer in which the repetitive information is concentrated into several nodes. If a certain node in a certain hidden layer contains more repetitive information, then since this node cannot learn more useful information, during the iterative update process of the weights, the change in the weight value of this node is relatively small compared to other nodes.

[0065] Preferably, in an embodiment of the present invention, the method for obtaining a node group in a target hidden layer includes:

[0066] Preset the first proportion threshold to 0.7, the second proportion threshold to 0.3, and the judgment threshold to 0.3, and take this as an example for description.

[0067] Taking any node A in any hidden layer of the neural network architecture as an example, obtain the modulus of each weight vector in the weight vector sequence of A to form a weight vector modulus sequence. Use the median in the weight vector modulus sequence as the reference weight value, obtain the number S of weight vector moduli in the weight vector modulus sequence that are less than the reference weight value, and then obtain the ratio of this number S to the length of the weight vector modulus sequence. When this ratio is greater than the preset first proportion threshold, record A as a target node.

[0068] It should be noted that: the obtaining of the modulus of a vector and the median in a sequence are both well-known techniques. The median of a data sequence refers to the value in the middle position after arranging a set of data in ascending or descending order. Thus, several target nodes in the hidden layer are obtained.

[0069] For any target node B, in the weight vector modulus sequence of B, select any two adjacent weight vector moduli C and D, and use the ratio of the absolute value of the difference between C and D to the maximum value of C and D as the difference value between C and D. When the maximum value among the difference values of all adjacent two weight vector moduli is less than the preset judgment threshold, record B as a repetitive information node.

[0070] In any hidden layer, when the ratio of the number of repetitive information nodes to the number of nodes in this any hidden layer is less than the preset second proportion threshold, record this any hidden layer as the target hidden layer.

[0071] Within any target hidden layer, obtain the inverse value of the cosine similarity between the scalar value sequences of any two nodes as the clustering distance between the any two nodes, and use the density clustering algorithm to perform clustering operations on all nodes to obtain several clustering clusters, and use each clustering cluster as each node group.

[0072] It should be noted that: both the density clustering algorithm and the cosine similarity are well-known technologies. The value range of the cosine similarity is between -1 and 1, and the larger the value, the more similar the two data sequences are. Therefore, in this embodiment, the difference between 1 and the cosine similarity between the scalar value sequences of any two nodes is used as the inverse value of the cosine similarity between the scalar value sequences of the any two nodes. It can be seen that the scalar value sequences of the nodes in each clustering cluster are similar.

[0073] Step S003: Delete the nodes in the node group to obtain an updated node group.

[0074] It should be noted that: if the nodes belonging to the same category in each hidden layer are directly grouped and then grouped convolution is performed, since there is also similar information in different pain data, this duplicate information needs to be removed, so as to obtain the similar information caused by the influence of the pain spread on the electrical signals at different parts of the urethra, that is, duplicate information. In view of this situation, in this embodiment, for each node in each node group in the hidden layer that needs to be grouped by nodes, the corresponding weight vector sequence is obtained, and then, through the larger values in the weight vector sequences of those nodes, it is determined whether the inputs that contribute more to these larger values belong to the same patient. If they belong to the same patient, then the node group obtained does not contain other pain information (the similar information existing in different pain data).

[0075] Preferably, in an embodiment of the present invention, the method for obtaining the updated node group includes:

[0076] For any node F in any node group of any target hidden layer, in the weight vector modulus sequence of F, obtain the mean value of all weight vector moduli, and count the ordinal values of all weight vector moduli greater than the mean value to form a weight ordinal value set.

[0077] It should be noted that: the weight ordinal value set corresponds to the larger weights of F during the weight iteration process, and can represent the weight change of this node well. Each weight ordinal value corresponds to a piece of data in the test dataset, that is, corresponds to a historical patient. If there are nodes in the node group to which this node belongs that are the same as the corresponding patient of this node, then the duplicate information carried by these same nodes does not contain other pain information (the similar information existing in different pain data), which is the similar information caused by the influence of the pain spread on the electrical signals at different parts of the urethra that this embodiment wants, that is, duplicate information.

[0078] Obtain all the data in the test dataset and input each piece of data into the neural network one by one. The input serial numbers of each piece of data are counted, and the set of input serial numbers of all the data of each historical patient in the test dataset is used as the input serial number set of each historical patient.

[0079] It should be noted that in this embodiment, the input serial numbers of each piece of data input into the neural network one by one from all the data in the test dataset are successively {1, 2, 3,...}.

[0080] If the set of weight order values of F is a subset of the input serial number set of any historical patient, then F is denoted as an initial retained node, and this arbitrary historical patient is used as the historical patient corresponding to the initial retained node F.

[0081] It should be noted that: for each initial retained node, all the data of the input corresponding to the larger weight are of the same historical patient.

[0082] In any node group, the historical patients corresponding to all the initial retained nodes form a historical patient set. The initial retained node corresponding to the historical patient with the largest number of the same historical patients in the historical patient set is denoted as the final retained node.

[0083] In any node group, all the final retained nodes form an updated node group.

[0084] Thus, the node deletion operation in the node group is completed.

[0085] It should be noted that: if the historical patient set is {Historical Patient 1, Historical Patient 1, Historical Patient 2, Historical Patient 3}, then the three initial retained nodes corresponding to Historical Patient 1 are used as the final retained nodes. If there are multiple historical patients with the largest number of the same historical patients in the historical patient set, any one of them is taken as an example for mechanical energy analysis. At this time, for each updated node group, the data corresponding to the larger weights of all the nodes in the neural network iteration process belong to the same historical patient, thereby avoiding other pain information in this node group, such as: similar information existing in the pain data of different patients.

[0086] Step S004: In the neural network architecture, each updated node group of each target hidden layer is regarded as a node to obtain the neural network after the architecture is updated; use the training dataset to train the neural network after the structure is updated to obtain the trained neural network; use the trained neural network to obtain the pain points during the catheter insertion process of the current patient and perform local anesthesia on the pain points.

[0087] Preferably, in an embodiment of the present invention, the method for obtaining pain points includes:

[0088] In the neural network architecture, each update node group of each target hidden layer is regarded as a node, and the neural network after updating the architecture is obtained.

[0089] It should be noted that in a neural network, the practice of treating multiple nodes in the same hidden layer as one node is usually referred to as "node aggregation" or "feature fusion", which is a well-known technology, and the specific method will not be introduced here. The schematic diagram of the neural network after updating the architecture is shown as Figure 4 shown. Figure 4 In the third column from left to right in [figure], the target hidden layer is shown, where the three boxed nodes form an update node group. All nodes in each update node group share a weight vector sequence, that is, the finally determined updated architecture is obtained. The network structure at this time is beneficial to removing similar information caused by different pain data, and then obtaining the similar information caused by the influence of the spread of pain on the electrical signals at different parts of the urethra, that is, duplicate information. This duplicate information is unified, avoiding the influence of redundant information caused by duplicate information on the network recognition function.

[0090] The training data set is input into the neural network after updating the architecture, and the mean squared error loss function is used as the loss function. Through iterative training, the trained neural network is obtained, which is a well-known operation.

[0091] According to the acquisition method of each piece of data of the above historical patients, each piece of data of the current patient during the catheter insertion process is acquired, and all the data of the current patient during the catheter insertion process are input into the trained neural network, and the pain points during the catheter insertion process of the current patient are output. Subsequently, local anesthesia is performed on the pain points.

[0092] It should be noted that after determining the location of the pain point, the permeable orifice of the catheter design will release lidocaine or other anesthetics as needed. The drug delivery process is carried out through the micropores or permeable orifices on the surface layer, and lidocaine will directly penetrate into the local area where the pain point is located, so as to achieve the effect of local anesthesia. This process can not only relieve the pain in the urethra, but also reduce the physical discomfort of the patient during the catheter insertion process.

[0093] The present invention also provides a urinary control management system based on an intelligent catheter, including a memory, a processor, and a computer program stored on the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the foregoing urinary control management method based on an intelligent catheter.

[0094] So far, the present invention is completed.

[0095] In summary, in the embodiments of the present invention, all historical data of all historical patients during the catheter insertion process are obtained to form a historical data set, and the historical data set is divided into a test data set and a training data set; all data in the test data set are input into the neural network one by one, and scalar value sequences and weight vector sequences of each node in each hidden layer in the neural network architecture are output; according to the magnitudes of the weights in the weight vector sequences of the nodes in each hidden layer, several target hidden layers are selected; in each target hidden layer, all nodes are divided into several node groups according to the similarity between the scalar value sequences of the nodes; the nodes in the node groups are deleted to obtain updated node groups; in the neural network architecture, each updated node group of each target hidden layer is used as a node to obtain the neural network after the architecture is updated; the updated neural network is trained using the training data set to obtain the trained neural network; the trained neural network is used to obtain the pain points during the catheter insertion process of the current patient, and local anesthesia is performed on the pain points. The present invention can accurately identify the pain points during the catheter insertion process of patients, and then perform local anesthesia on the pain points to improve the effect of urinary control management.

[0096] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A urine control management system based on an intelligent urinary catheter, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is executed by a processor, the following steps are implemented: Acquire a historical data set consisting of all pieces of data of all historical patients during the process of catheter insertion, wherein the data is collected through microelectrodes on the surface of the catheter, and one piece of data is formed by three voltages within every three seconds and three-dimensional coordinates of the last second; divide the historical data set into a test data set and a training data set; input all pieces of data in the test data set into the neural network one by one, and output a scalar value sequence and a weight vector sequence of each node in each hidden layer in the neural network architecture; According to the size of the modulus of the weight vector in the weight vector sequence of the nodes in each hidden layer, a number of target hidden layers are selected; in each target hidden layer, according to the similarity between the scalar value sequences of the nodes, all the nodes are divided into a number of node groups; Delete the nodes in the node group to obtain an updated node group; In the neural network architecture, each updated node group of each target hidden layer is regarded as a node to obtain a neural network with an updated architecture; the neural network with the updated structure is trained using a training data set to obtain a trained neural network; Use the trained neural network to obtain the pain points of the current patient during the catheter insertion process; The specific steps of deleting the nodes in the node group to obtain the updated node group are as follows: For any node in any node group of any target hidden layer, denoted as node F, the module of each weight vector in the weight vector sequence of node F constitutes the weight vector module sequence of node F; In the weight vector modulus sequence of node F, the mean of all weight vector moduli is obtained, and the ordinal values ​​of all weight vector moduli greater than the mean are counted to form a weight ordinal value set; Obtain the input sequence number of each data item in the test data set and input it into the neural network one by one, and count the input sequence number set of all data items in the test data set for each historical patient; According to the weight sequence value set and the input sequence number set of each historical patient, the initial reserved nodes are selected from any node group, and the historical patients corresponding to each initial reserved node are obtained; According to the historical patients corresponding to the initial reserved nodes, the final reserved nodes are screened out from the initial reserved nodes; All the final retained nodes in any node group form an update node group.

2. According to claim 1, a urine control management system based on an intelligent urinary catheter is characterized in that: The specific steps of screening out a number of target hidden layers are as follows: Obtain the modulus of each weight vector in the weight vector sequence of any node in any hidden layer to form a weight vector modulus sequence; Filter out the target node according to the size of the weight vector modulus in the weight vector modulus sequence of each node in the hidden layer; According to the difference of adjacent weight vector moduli in the weight vector modulus sequence of each target node, duplicate information nodes are screened out; According to the number of repeated information nodes in each hidden layer, the target hidden layer is selected.

3. According to claim 2, a urine control management system based on an intelligent urinary catheter is characterized in that: The specific steps of screening out the target node according to the size of the weight vector modulus in the weight vector modulus sequence of each node in the hidden layer include the following: Select any node A in any hidden layer, and take the median in the weight vector modulus sequence of A as the benchmark weight value; obtain the number S of weight vector moduli in the weight vector modulus sequence of A that are less than the benchmark weight value, and when the ratio of the number S to the length of the weight vector modulus sequence of A is greater than a preset first proportion threshold, record A as the target node.

4. According to claim 2, a urine control management system based on an intelligent urinary catheter is characterized in that: The specific steps of screening out duplicate information nodes according to the difference of adjacent weight vector modules in the weight vector module sequence of each target node are as follows: In the weight vector modulus sequence of any target node B, select any two adjacent weight vector moduli C and D, and take the ratio of the absolute value of the difference between C and D to the maximum value between C and D as the difference value of C and D. When the maximum value of the difference values ​​of all two adjacent weight vector moduli is less than the preset judgment threshold, B is recorded as a duplicate information node.

5. The urine control management system based on the intelligent urinary catheter according to claim 2, characterized in that: The method of selecting a target hidden layer according to the number of repeated information nodes in each hidden layer includes the following specific steps: In any hidden layer, when the ratio of the number of repeated information nodes to the number of nodes in the any hidden layer is less than a preset second proportion threshold, the any hidden layer is recorded as a target hidden layer.

6. According to claim 1, a urine control management system based on an intelligent urinary catheter is characterized in that: In each target hidden layer, all nodes are divided into several node groups according to the similarity between the scalar value sequences of the nodes, and the specific steps included are as follows: In any target hidden layer, an inverse proportional value of the cosine similarity between the scalar value sequences of any two nodes is obtained as the clustering distance of the any two nodes, and a clustering operation is performed on all nodes to obtain a plurality of node groups.

7. The urine control management system based on the intelligent urinary catheter according to claim 1, characterized in that: The specific steps of selecting the initial reserved nodes from any node group according to the weight sequence value set and the input sequence number set of each historical patient and obtaining the historical patients corresponding to each initial reserved node are as follows: If the weight sequence value set of any node in any node group is a subset of the input sequence number set of any historical patient, the arbitrary node is recorded as the initial reserved node, and the arbitrary historical patient is used as the historical patient corresponding to the initial reserved node.

8. The urine control management system based on the intelligent urinary catheter according to claim 1, characterized in that: The method of selecting the final reserved nodes from the initial reserved nodes according to the historical patients corresponding to the initial reserved nodes includes the following specific steps: In any node group, the historical patients corresponding to all the initial reserved nodes constitute a historical patient set, and the initial reserved node corresponding to the historical patient with the largest number of the same historical patients in the historical patient set is recorded as the final reserved node.

Images